DiDeMo
Emerging11papers using it
2021first seen
About DiDeMo contains 10K long-form videos from Flickr. For each video, ~4 short sentences are annotated in temporal order. We follow the existing works to concatenate those short sentences and evaluate ‘paragraph-to-video’ retrieval on this benchmark. We adopt the official split: Train: 8,395 videos, 8,395 captions (c
Papers using DiDeMo (11)
- InstAP: Instance-Aware Vision-Language Pre-Train for Spatial-Temporal UnderstandingBima: Towards Biases Mitigation For Text-video Retrieval Via Scene Element GuidanceDelving Deeper: Hierarchical Visual Perception for Robust Video-Text RetrievalFrom Captions To Keyframes: Keyscore For Multimodal Frame Scoring And Video-language UnderstandingGAIS: Frame-level Gated Audio-visual Integration With Semantic Variance-scaled Perturbation For Text-video RetrievalWeakly Supervised Temporal Adjacent Network for Language GroundingCLIP-ViP: Adapting Pre-trained Image-Text Model to Video-Language
Representation AlignmentCross-Modal Adapter for Vision-Language RetrievalMuMUR : Multilingual Multimodal Universal RetrievalTagging before Alignment: Integrating Multi-Modal Tags for Video-Text
RetrievalMug-STAN: Adapting Image-Language Pretrained Models for General Video
Understanding